We have shown that administration of inorganic potassium phosphates (Pi) to patients with severe diabetic ketoacidosis was able to increase the P50 (the PO2 necessary to achieve a hemoglobin saturation of 50%) by a non diphosphoglycerate (DPG) mediated effect. This suggests that the oxyhemoglobin dissociation curve (ODC) may be determined not only by pH, temperature, CO2 content and DPG but also by plasmatic ions. In order to test this hypothesis we have determined the ODC on whole blood in two groups of subjects, 49 control subjects with matching age and sex and 49 patients suffering from liver cirrhosis, acute pancreatitis, septic shock and acute respiratory distress syndrome. The patients had many ionic disorders induced either by their diseases or by the applied treatment. The mean ODC of the patients did not differ from the normal values. In contrast, the dispersion of PO2 around the saturations values was increased from 5 to 80% saturation. A forward regression analysis showed that the DPG level and the levels of inorganic phosphates and natrium (Na+) played a significant role in determining the position of the ODC according to the following equation: P50 (mmHg) = 34.5 + 0.225 DPG + 0.62 Pi-0.09 Na+, where DPG is in micromol.gHb-1 and Pi and Na+ in mEq.l-1. In separate experiments we showed that the Bohr effect as expressed in d (log PO2)/dpH amounted to -0.53, -0.46 and -0.42 for SO2 equal to 5%, 50% and 95%, respectively. The corresponding values for the temperature effect was expressed in d (log PO2)/dT amounted to 0.028, 0.024, and 0.020 respectively. The fact that ions play an role in regulating the position of the ODC of patients with ionic disorders may have therapeutical implications, preventive or curative.